The Agent Skills Directory

[NO_CODE]: The skill consists entirely of natural language design guidelines and does not contain any executable scripts, shell commands, or binary components.\n- [PROMPT_INJECTION]: The instructions are constrained to design methodology and brand alignment; while the skill ingests context from the current codebase (ingestion point), it lacks any functional capabilities like file writing or network access that would make indirect injection exploitable.\n- [DATA_EXFILTRATION]: No network operations or commands to read sensitive files (such as credentials or private keys) were identified in the skill logic.\n- [REMOTE_CODE_EXECUTION]: The skill does not download external packages (npm/pip) or execute remote scripts; it only references a separate design skill as a conceptual dependency.\n- [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or other sensitive authentication secrets were detected in the skill's content or metadata.

colorize